A paper published this month stopped me mid-morning-tea. The title alone: “The body does not keep the score: trauma, predictive coding, and the restoration of metastability.”
Published in Frontiers in Systems Neuroscience by Steven Kotler, Michael Mannino, Glenn Fox, and Karl Friston — that last name carries serious weight in computational neuroscience — this piece offers something I’ve been waiting for: a theoretical framework that validates body-based recovery work while correcting a metaphor they feel have gone too far.
The short version
Van der Kolk’s The Body Keeps the Score shaped how an entire generation thinks about trauma. The book’s contributions are real. But the metaphor — that trauma is literally stored in bodily tissue — has been taken at face value in ways that can mislead both practitioners and clients.
Kotler and colleagues argue that trauma isn’t stored anywhere. It’s reenacted — dynamically, in real time — by a brain stuck in rigid threat predictions. Your brain over-weights danger signals, interprets its own arousal (racing heart, muscle tension, shallow breathing) as evidence that the threat is still present, and locks into a self-confirming loop. The “score” the body appears to keep is an artifact of circular inference: the brain predicts pain, senses arousal, and takes that arousal as proof.
The body participates in trauma as messenger, not archive.
What actually needs to heal
The authors frame the goal of recovery as restoring metastability — the brain’s capacity to fluidly shift between different network states depending on context. In a healthy system, neural networks assemble, collaborate, and dissolve as needed. Flexible, responsive, adaptive.
In PTSD, that fluidity collapses. The brain gets trapped in narrow defensive configurations. Connectivity patterns shift toward threat detection. The system’s capacity for flexible reconfiguration shrinks.
Recovery, then, isn’t about finding and purging what was buried. It’s about restoring the brain’s ability to move.
Why body-based work actually works (according to this model)
Here’s where it gets interesting for those of us doing somatic, movement-based, and sensory regulation work. The paper explicitly addresses why these approaches are effective — and the answer isn’t “releasing stored trauma.”
Body-based interventions work because they introduce novel sensorimotor feedback that recalibrates the brain’s predictions. When you engage in somatic movement, breathwork, or sensory grounding, you’re giving the nervous system new data that doesn’t match its threat expectations. Over time, this expands the brain’s repertoire of viable states. The system relearns that arousal and safety can coexist.
The authors also highlight flow states — deep absorption in meaningful, challenging activity — as a particularly potent mechanism. During flow, brain networks rapidly reconfigure in adaptive ways. The prefrontal cortex shifts its processing. Reward and attention systems engage. The authors hypothesize that flow restores the very metastable dynamics that trauma disrupts. (They cite surf therapy and hiking programs for active-duty military as early evidence, though they’re careful to note that the specific contribution of flow can’t yet be isolated from other factors like exercise and social connection.)
A framework that unifies, not divides
One of the most useful contributions of this paper: it explains why so many different treatments can all work for trauma. EMDR, mindfulness, exposure therapy, exercise, psychedelics, flow-inducing activities — they look different on the surface. But in this framework, they all accomplish the same thing at the network level: restoring flexible coupling between large-scale brain networks, quieting maladaptive self-referential loops, and rebalancing neuromodulation.
The mechanism isn’t specific content. It’s dynamic reorganization.
That’s a clarifying idea. It dissolves turf wars between modalities and points toward a shared underlying principle.
What this framework is — and isn’t
Honesty matters here. This is an opinion paper presenting a theoretical framework, not an empirical study. The core claim — that PTSD specifically involves reduced metastability — hasn’t been directly tested using established metrics in PTSD populations. The metastability research the authors cite comes from traumatic brain injury studies, which is a neurologically distinct condition. The authors are transparent about this gap and call for future research.
The paper also doesn’t dismiss van der Kolk. It acknowledges his substantial contributions while arguing that the metaphor has outrun the science in popular reception — reinforcing the mind-body duality that his work aimed to dissolve.
What I take from this
As someone who works at the intersection of somatic embodiment, nervous system regulation, and metabolic health, this framework resonates. It gives us more precise language for what we’re already doing. It doesn’t change the practice. It sharpens how we understand and communicate it.
The line that will stay with me: Healing is not excavation but exploration.
Your body isn’t keeping score. Your brain is running an outdated prediction. And the work — our work — is helping it update.
Full paper: Kotler, Mannino, Fox & Friston (2026). “The body does not keep the score: trauma, predictive coding, and the restoration of metastability.” Frontiers in Systems Neuroscience, 20. Read it here.